Automatic detection of corrosion and associated damages to civil infrastructures such as bridges, buildings, and roads, from aerial images captured by an Unmanned Aerial Vehicle (UAV), helps one to overcome the challenges and shortcomings (objectivity and reliability) associated with the manual inspection methods. Deep learning methods have been widely reported in the literature for civil infrastructure corrosion detection. Among them, convolutional neural networks (CNNs) display promising applicability for the automatic detection of image features less affected by image noises. Therefore, in the current study, we propose a modified version of deep hierarchical CNN architecture, based on 16 convolution layers and cycle generative adversarial network (CycleGAN), to predict pixel-wise segmentation in an end-to-end manner using the images of Bolte Bridge and sky rail areas in Victoria (Melbourne). The convolutedly designed model network proposed in the study is based on learning and aggregation of multi-scale and multilevel features while moving from the low convolutional layers to the high-level layers, thus reducing the consistency loss in images due to the inclusion of CycleGAN. The standard approaches only use the last convolutional layer, but our proposed architecture differs from these approaches and uses multiple layers. Moreover, we have used guided filtering and Conditional Random Fields (CRFs) methods to refine the prediction results. Additionally, the effectiveness of the proposed architecture was assessed using benchmarking data of 600 images of civil infrastructure. Overall, the results show that the deep hierarchical CNN architecture based on 16 convolution layers produced advanced performances when evaluated for different methods, including the baseline, PSPNet, DeepLab, and SegNet. Overall, the extended method displayed the Global Accuracy (GA); Class Average Accuracy (CAC); mean Intersection Of the Union (IOU); Precision (P); Recall (R); and F-score values of 0.989, 0.931, 0.878, 0.849, 0.818 and 0.833, respectively.

In the evolving landscape of consumer electronics, the Generative Adversarial Learning-based Trust Management (GALTrust) framework emerges as a novel solution, uniquely combining Generative Adversarial Networks (GANs) and type-2 fuzzy logic to tackle trust management challenges within the Internet of Drone Things (IoDT). Addressing the pivotal needs of spatial crowdsourcing scenarios like bushfire management, GALTrust significantly overcomes the limitations posed by traditional machine learning methods in detecting emergent types of malicious nodes and navigating the impact of training data size variations. At its core, GALTrust features a GAN-based codec structure, meticulously trained with trust vectors, enabling precise differentiation between malicious and trustworthy nodes. A key innovation of GALTrust is the introduction of a GAN-based trust redemption model, strategically designed to curtail false positives and safeguard against the unwarranted exclusion of benign drones, thus markedly enhancing network resilience. This framework exhibits dynamic adaptability, continually refining its trust model to align with the latest detection insights within the IoDT ecosystem. Through its application in secure clustering for IoDT, GALTrust has proven its efficacy by achieving an exceptional detection rate of up to 94.1% and maintaining a false positive rate below 9.1%, thereby significantly elevating security and operational efficiency in crucial consumer electronics applications.

In this study, we introduce a pioneering framework, DroneSSL, that integrates the concept of spatial crowdsourcing with TinyML to enhance anomaly detection in the Internet of Drone Things (IoDT). This innovative approach leverages drones and unmanned ground vehicles (UGVs) for expansive data collection in environments that are typically inaccessible or hazardous, such as during Australian bushfire incidents. By employing lightweight machine learning models alongside advanced communication technologies, DroneSSL transcends traditional spatial-temporal data analysis methods. It efficiently processes multimodal data from diverse Points-of-Interest (PoIs), significantly improving the quality and speed of data collection and analysis. The framework’s integration of a temporal feature extraction module with a Graph Neural Network (GNN) and its adaptable, scalable GNN architecture tailor DroneSSL for real-time operations in resource-constrained IoDT environments. Achieving an 89.6% F1 score, DroneSSL marks a substantial 4.9% improvement over existing approaches, highlighting its effectiveness in critical applications such as environmental surveillance and emergency response. This advancement not only showcases the potential of combining TinyML with spatial crowdsourcing for IoDT but also sets a new standard for efficient, scalable anomaly detection, paving the way for future innovations in IoT edge devices and environmental monitoring systems.

We introduce the AI-Generated Optimal Decision (AIGOD) algorithm and the Deep Diffusion Soft Actor-Critic (DDSAC) framework, marking a significant advancement in integrating Human Digital Twins (HDTs) with AI-Generated Content (AIGC) within IoMT-based smart homes. Our innovative AI-Generated Content-as-a-Service (AIGCaaS) architecture, optimized for IoMT environments, leverages network edge servers to enhance the selection of AI-Generated Content Service Providers (AISPs) tailored to the unique characteristics of individual HDTs. Extensive experiments demonstrate DDSAC’s HDT-centric approach outperforms traditional Deep Reinforcement Learning algorithms, offering optimal AIGC services for diverse healthcare needs. Specifically, DDSAC achieved a 20% improvement in task completion rates and a 15% increase in overall utility compared to existing methods. These findings highlight the potential of HDTs in personalized healthcare by simulating and predicting patient-specific medical outcomes, leading to proactive and timely interventions. This integration facilitates personalized healthcare, establishing a new standard for patient-centric care in smart home environments. By leveraging cutting-edge AI techniques, our research significantly contributes to the fields of IoMT and AIGC, paving the way for smarter and more responsive healthcare services.

We introduce an ensemble model approach for multimodal sentiment analysis, focusing on the fusion of textual and video data to enhance the accuracy and depth of emotion interpretation. By integrating three foundational models-IFFSA, BFSA, and TBJE-using advanced ensemble techniques, we achieve a significant improvement in sentiment analysis performance across diverse datasets, including MOSI and MOSEI. Specifically, we propose two novel models-IFFSA and BFSA, which utilise the large language models BERT and GPT-2 to extract the features from text modality and ResNet and VGG for video modality. Our work uniquely contributes to the field by demonstrating the synergistic potential of combining different modal analytical strengths, thereby addressing the intricate challenge of nuanced emotion detection in multimodal contexts. Through comprehensive experiments and an extensive ablation study, we not only validate the superior performance of our ensemble model against current state-of-the-art benchmarks but also reveal critical insights into the model's capability to discern complex emotional states. Our findings underscore the strategic advantage of ensemble methods in multimodal sentiment analysis and set a new precedent for future research in effectively integrating multimodal data sources.

We introduce “TMIoDT,” a pioneering framework aimed at bolstering communication security in the Internet of Drone Things (IoDT) integrated with Open Radio Access Networks (Open RAN), with a specific focus on bushfire monitoring applications. Our novel contributions include the seamless integration of digital twin technology with blockchain to establish a robust trust management system in the IoDT context. This approach addresses the critical vulnerabilities associated with unsecured wireless networks in IoDT, such as data integrity issues and susceptibility to cyber threats. The TMIoDT framework encompasses a mutual authentication mechanism to secure interactions and key exchanges among IoDT entities, including drones and Unmanned Ground Vehicles (UGVs). Furthermore, it leverages blockchain technology for credible trust management and employs digital twins to model UGV servers accurately, enhancing IoDT relationship modeling. An advanced Intrusion Detection System (IDS), utilizing Stacked Variational Autoencoder (SVA) and Attention-based Bidirectional LSTM (ABL), is implemented for anomaly detection, complemented by a blockchain-based transaction writing scheme for secure data verification. Our comprehensive evaluation, utilizing the ToN-IoT and ICIDS-2017 network intrusion datasets, confirms TMIoDT’s effectiveness in significantly improving communication security and reliability in IoDT.